News

Algorithm developed that can predict gestational diabetes

A computer algorithm has been created that can predict gestational diabetes.

A team from the Weizmann Institute of Science, Israel, developed an algorithm that can identify women at high risk of developing gestational diabetes. The prediction can be reliably made in the early stages of pregnancy, or even before pregnancy.

Gestational diabetes is typically diagnosed during the third trimester (between 24 and 28 weeks of pregnancy). It is characterised by high blood glucose levels which only develop during pregnancy. In total, between 3 and 9 per cent of pregnant women develop the condition. It can cause health complications for mothers and babies during pregnancy and after birth.

The algorithm was tested on electronic health records of almost 600,000 pregnancies obtained from Clalit Health Services.

The researchers applied a machine learning method to the data on 450,000 of the pregnancies, which occurred from 2010 to 2017. Gestational diabetes had been established in about 4 per cent of these cases. A total of 2,000 parameters were measured to see how sufficient they were to predict gestational diabetes. These parameters included blood test results and family histories.

According to the results, these 2,000 parameters were whittled down to just nine that were found to be most effective and were sufficient to correctly identify those with high risk of developing the condition. These nine factors included age, body mass index and family history of diabetes, and blood glucose tests taken during a subject’s previous pregnancies.

In order to test the algorithm, it was applied to 140,000 pregnancy records that had not been used in the initial analysis. The findings showed the algorithm was able to accurately identify the women who developed gestational diabetes, which validated the findings from the study. It could be possible to establish in advance whether a woman is at a high risk of developing gestational diabetes by answering just nine questions, according to researchers.

The ability to identify the women most at risk of developing gestational diabetes would enable healthcare professionals to intervene earlier. It could lead to better support and encouragement for pregnant women to undertake appropriate lifestyle changes that could reduce the risk or mitigate the blood glucose spikes of gestational diabetes.

Professor Eran Segal helped to lead the study. He is from the Institute’s Department of Computer Science and Applied Mathematics, and the Department of Molecular Cell Biology. He said: “Our ultimate goal has been to help the health system take measures so as to prevent diabetes from occurring in pregnancy.”

The research was published in the journal Nature Medicine.

To Top